T2K Fusion Framework
SNAPSHOT 20160525
© 2016 University of
Mannheim, Oliver Lehmberg, Yaser Oulabi and Dominique Ritze
Please see LICENSE.txt
The purpose of this code is to provide an early insight into the fusion
process presented in a scientific research paper. This code is
not suitable for productive use as it is a snapshot of a
software in development. The code is not commented nor is any
documentation provided.
Table of Contents
- Requirements
- Building the application
- Preparing the data folder
- Running the application
Requirements
Technical requirements:
- 280 GB of RAM Memory
- JDK 8
- Maven
Data requirements
Included data requirements
Building the application
You first need to build the T2K Fusion Framework, which includes
compiling the code and resolving dependencis.
Building the project is only supported through maven.
Run the command mvn in the src
folder provided with this archive. This should successfuly compile the
sources, resolve all dependencies and package the application.
This will create the following runnable JAR file:
src/target/expansion-0.0.1-SNAPSHOT-jar-with-dependencies.jar
Preparing the data folder
The data directory contains all necessary input data. It is also the
location where all the output is stored.
The data directory must be structured as following. Use the baseDataDir
directory provided in this package as a starting point.
dataDir
|- tables/
|- cache/
|- columnMappings/
|- mappings/
|- rankings/
|- results/
-
|- rankings/
|- heidelTime/
|- Units/
|- Convertible/
|- split.config
|- empty.txt
|- dbpedia_2014.owl
-
After copying the empty data directory from the project you need to
undertake the following tasks:
- Extract the Table Mappings to the directory
dataDir/tables/mappings/
- Extract the Column Mappings files to the directory
dataDir/tables/columnMappings/
- Extract the PageRank ranking file to the directory
dataDir/tables/rankings/
Running the application
When running the application you need to
- use Java 8
- set the Java Maximum Memory Paramter to at least 280 GB
- use the Java G1 Garbage Collector
- add the location of the dataDir to the first argument
passed to the Java Runtie Environemnt
Example run of the application
jre/bin/java -Xmx280000M -XX:+UseG1GC -jar expansion-0.0.1-SNAPSHOT-jar-with-dependencies.jar "/home/user/t2k/dataDir/"
After starting the application you will be presented with multiple choiches in the console.